- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Tue, 4 Oct 2011 16:04:14 -0700
- To: Brian Manthos <brianman@microsoft.com>
- Cc: "L. David Baron" <dbaron@dbaron.org>, Boris Zbarsky <bzbarsky@mit.edu>, Sylvain Galineau <sylvaing@microsoft.com>, "www-style@w3.org" <www-style@w3.org>
On Tue, Oct 4, 2011 at 1:54 PM, Brian Manthos <brianman@microsoft.com> wrote: > Tab: >> If you ignore >> performance optimizations, the only thing that display:none really >> does is prevent you from doing "used values" > > I realize it's an old draft and the language has changed in the CSS4 flavor, but cases like this are part of what I was concerned about: > http://www.w3.org/TR/2011/WD-css3-images-20110217/#interpolating-gradients > 11.2, second list > # 1.Convert both the start and end gradients to their used value. Yes, internal argument over how gradient interpolation should work is part of what made me remove it for now. ^_^ I now think I agree with Simon Fraser's stance that transitions definitely take place over computed values; that implies that we need to invent a syntax to represent computed-level intermediate states, for the times when you can't literally use the computed values (something like "transition(<val1>,<val2>,<percentage>)"). That way, you can still carry around a computed value that just represents the blend, and then resolve it down to the exact value later, similar to transitioning from "20px" to "50%" using calc(). > I'm also curious what "display:none;" does to percentage values (everywhere) that are resolved against the used width/height w/r/t transitioning. I'm not going to try and hunt it down right now, but I *think* 2.1 already resolves this by saying they resolve to 0. If not, some language like that can be minted. It's a solvable problem. ~TJ
Received on Tuesday, 4 October 2011 23:05:02 UTC